So, where are the robots? We've been told for 40 years already that they're coming soon. Very soon they'll be doing everything for us. They'll be cooking, cleaning, buying things, shopping, building. But they aren't here. Meanwhile, we have illegal immigrants doing all the work, but we don't have any robots. So what can we do about that? What can we say? So I want to give a little bit of a different perspective of how we can perhaps look at these things in a little bit of a different way. And this is an x-ray picture of a real beetle, and a Swiss watch, back from '88. You look at that -- what was true then is certainly true today. We can still make the pieces. We can make the right pieces. We can make the circuitry of the right computational power, but we can't actually put them together to make something that will actually work and be as adaptive as these systems.
Deci, unde sunt robotii? De 40 de ani ni se spune ca vor veni in curand. In curand o sa faca totul in locul nostru: o sa gateasca,o sa faca curat,o sa faca cumparaturi, o sa construiasca.Dar nu sunt aici. Intre timp ,avem imigranti ilegali care fac toata treaba, dar nu avem nici un robot. Asa ca ce putem sa facem? Ce putem sa spunem? As vrea sa va arat o modalitate alternativa despre cum ne putem uita la lucruri intr-un mod putin diferit. Aceasta este o radiografie a unui gandac adevarat, si a unui ceas elvetian, din '88.Te uiti la -- ce era adevarat atunci cu siguranta este adevarat si astazi. Inca mai putem sa fabricam piesele, putem face piesele corecte. putem sa facem o placuta cu circuite de calcul, dar nu putem sa le punem la un loc sa facem ceva anume care sa functioneze cu adevarat si sa fie capabil sa se adapteze la fel ca aceste sisteme.
So let's try to look at it from a different perspective. Let's summon the best designer, the mother of all designers. Let's see what evolution can do for us. So we threw in -- we created a primordial soup with lots of pieces of robots -- with bars, with motors, with neurons. Put them all together, and put all this under kind of natural selection, under mutation, and rewarded things for how well they can move forward. A very simple task, and it's interesting to see what kind of things came out of that.
Asa ca sa incercam sa privim lucrurile dintr-o alta perspectiva. Sa-l luam pe cel mai bun designer, cel mai bun designer dintre toti: sa vedem ce poate face evolutia pentru noi. Asa ca am amestecat-- am creat supa primordiala cu multe bucati de roboti : cu fiare, cu motoare , cu neuroni. Le adunam pe toate la un loc, si le supunem unui fel de proces natural de selectie, unui proces de transformare, si vedem cat de bine au sa evolueze. O sarcina foarte simpla, si e interesant de vazut ce fel de chestii rezulta.
So if you look, you can see a lot of different machines come out of this. They all move around. They all crawl in different ways, and you can see on the right, that we actually made a couple of these things, and they work in reality. These are not very fantastic robots, but they evolved to do exactly what we reward them for:
Asa ca daca va uitati, o sa vedeti o gramada de masinarii diferite care au iesit din asta.Toate se misca, intr-un fel sau altul,puteti vedea in dreapta, chiar am creat niste chestii de astea, si chiar functioneaza. Nu sunt cine stie ce roboti, dar au evoluat si au ajuns sa facea ce le-am cerut:
for moving forward. So that was all done in simulation, but we can also do that on a real machine. Here's a physical robot that we actually have a population of brains, competing, or evolving on the machine. It's like a rodeo show. They all get a ride on the machine, and they get rewarded for how fast or how far they can make the machine move forward. And you can see these robots are not ready to take over the world yet, but they gradually learn how to move forward, and they do this autonomously.
se mearga inainte.Toate aceastea au fost facute intr-o simulare, dar putem face asta si cu o masinarie reala. Acesta este un robot pe care avem o populatie de creiere, care concureaza unele cu celelalte, sau evoluaza, pe robot. E ca la un rodeo show: toti apuca sa controleze masinaria, si sunt recompensati pentru cat de repede sau cat de departe au facut masinaria sa mearga. Dupa cum vedeti acesti roboti nu sunt gata inca sa preia controlul asupra lumii,dar invata treptat cum sa se miste inainte, si fac aceste lucru in mod autonom.
So in these two examples, we had basically machines that learned how to walk in simulation, and also machines that learned how to walk in reality. But I want to show you a different approach, and this is this robot over here, which has four legs. It has eight motors, four on the knees and four on the hip. It has also two tilt sensors that tell the machine which way it's tilting.
Deci in aceste doua exemple, am avut de fapt masinarii care au invatat cum sa mearga intr-o simulare, si masinarii care au invatat sa mearga in realitate. Dar vreau sa va arat o abordare diferita, si acesta este robotul, aici, care are patru picioare, are opt motoare , patru la genunchi si patru la solduri. Mai are si doi senzori care ii spun masinariei in ce parte sa se incline.
But this machine doesn't know what it looks like. You look at it and you see it has four legs, the machine doesn't know if it's a snake, if it's a tree, it doesn't have any idea what it looks like, but it's going to try to find that out. Initially, it does some random motion, and then it tries to figure out what it might look like. And you're seeing a lot of things passing through its minds, a lot of self-models that try to explain the relationship between actuation and sensing. It then tries to do a second action that creates the most disagreement among predictions of these alternative models, like a scientist in a lab. Then it does that and tries to explain that, and prune out its self-models.
Dar aceasta masinarie nu stie cum arata. Tu te uiti la ea si vezi ca are patru picioare, masinaria nu stie daca e un sarpe, daca e un copac, nu are nici o idee despre cum arata, dar o sa incerce sa afle. Initial, o sa incerce niste miscari aleatorii, si apoi incearca sa afle cum arata -- si vedeti cum o gramada de lucruri ii trec prin minte, o gramada de auto-modele care incearca sa explice relatia dintre actiune si raspuns-- si apoi incearca o a doua actiune care creaza dezacord printre predictiile modelelor alternative, ca un om de stiinta intr-un laborator. Apoi face asta si incearca sa explice, si sa isi intreaca concurentii.
This is the last cycle, and you can see it's pretty much figured out what its self looks like. And once it has a self-model, it can use that to derive a pattern of locomotion. So what you're seeing here are a couple of machines -- a pattern of locomotion. We were hoping that it wass going to have a kind of evil, spidery walk, but instead it created this pretty lame way of moving forward.
Acesta e ultimul ciclu, si dupa cum puteti vedea si-a dat seama cum arata,odata ce a avut un model dupa care sa se ia, se poate lua dupa asta ca sa isi creeze un tipar de locomotie. Deci ce vedeti aici este o adunatura de masinarii-- un tipar de locomotie. Speram ca o sa aiba un mers "smecher" ,ca al unui paianjen, dar in schimb,si-a creat acest mod nasol de a se misca inspre inainte.
But when you look at that, you have to remember that this machine did not do any physical trials on how to move forward, nor did it have a model of itself. It kind of figured out what it looks like, and how to move forward, and then actually tried that out. (Applause)
Dar cand te uiti la asta , trebuie sa tii cont ca aceasta masinarie nu stia cum sa se miste inainte, nici nu stia cum arata. Si-a dat seama cum arata , si cum sa se miste, si apoi a facut o incercare. (Aplauze)
So, we'll move forward to a different idea. So that was what happened when we had a couple of -- that's what happened when you had a couple of -- OK, OK, OK -- (Laughter) -- they don't like each other. So there's a different robot. That's what happened when the robots actually are rewarded for doing something. What happens if you don't reward them for anything, you just throw them in?
Asa, se ne indreptam atentia spre o idee diferita. Deci asta sa intamplat cand am avut o gramada de -- asta sa intamplat cand am avut o gramada de -- Ok ,Ok ,Ok-- (Rasete) -- nu se plac.Deci e un robot diferit. Asta sa intamplat cand robotii au fost recompensati pentru ca fac ceva. Dar ce se intampla cand nu ii recompensezi, doar ii arunci acolo?
So we have these cubes, like the diagram showed here. The cube can swivel, or flip on its side, and we just throw 1,000 of these cubes into a soup -- this is in simulation --and don't reward them for anything, we just let them flip. We pump energy into this and see what happens in a couple of mutations. So, initially nothing happens, they're just flipping around there. But after a very short while, you can see these blue things on the right there begin to take over.
Deci avem cuburile astea,dupa cum arata diagrama asta. Cubul poate sa pivoteze ,sau sa sara pe o parte, si aruncam 1,000 de cuburi de astea intr-o supa-- asta intr-o simulare-- si nu ii recompensam pentru nimic. ii lasam acolo sa sara. Le dam energie si vedem ce se intampla in cateva mutatii. Initial, nimic nu se intampla, doar sar de colo colo. Dar dupa o scurta perioada de timp,puteti vedea aceste chestii albastre din dreapta incep sa preia controlul.
They begin to self-replicate. So in absence of any reward, the intrinsic reward is self-replication. And we've actually built a couple of these, and this is part of a larger robot made out of these cubes. It's an accelerated view, where you can see the robot actually carrying out some of its replication process. So you're feeding it with more material -- cubes in this case -- and more energy, and it can make another robot. So of course, this is a very crude machine, but we're working on a micro-scale version of these, and hopefully the cubes will be like a powder that you pour in.
Incep sa se auto-reproduca.Asa ca in absenta vreunei recompense, propria recompensa este auto-reproducerea. Si chiar am construit cativa din astia , si asta e o parte dintr-un robot mai mare facut din aceste cuburi, e o filmare accelerata, in care puteti vedea cum robotul urmeaza pasii spre procesul de replicare. Deci o hranim cu mai mult material-- cuburi in cazul de fata-- si mai multa energie, si poate face un alt robot. Dar desigur , aceasta este o masinarie foarte primitiva, dar lucram la versiuni microscopice ale acestor masinarii, si speram ca ,cuburile o sa fie ca o pudra pe care o adaugi.
OK, so what can we learn? These robots are of course not very useful in themselves, but they might teach us something about how we can build better robots, and perhaps how humans, animals, create self-models and learn. And one of the things that I think is important is that we have to get away from this idea of designing the machines manually, but actually let them evolve and learn, like children, and perhaps that's the way we'll get there. Thank you. (Applause)
OK, deci ce putem invatat? Acesti roboti nu sunt desigur foarte folositori, dar ne pot invata cate ceva despre cum putem sa contruim roboti mai buni, si poate cum oameni , animalele, pot crea auto-modele si invata. Si unul din lucrurile pe care il consider important este acela ca trebuie sa ne indepartam de idea de a proiecta manual masinariile, si in schimb sa le lasam sa evolueze si sa invete,precum copiii, si poate in felul acesta o sa reusim . Multumesc. (Aplauze)